Previous study using rapid T2 mapping techniques such as mcDESPOT have shown the feasibility of multi-component T2 analysis of human knee joint (1,2). A modified mcDESPOT termed as mcRISE was also proposed to correct the signal derivation raised by the magnetization transfer (MT) (3). Both of those techniques take advantages of steady-state sequences for rapid acquisition. However, recent studies have shown the steady-state signal could be affected by the finite pulse effect (4,5). Without proper modeling, the finite pulse effect may bias relaxation parameter estimation (4,5). In this study, we incorporated the finite pulse correction in the mcRISE model and compared it with original mcDESPOT and mcRISE without correction. The modified mcRISE model may provide both MT and finite pulse effect in-sensitive T2 parameters which can increase the robustness and accuracy for multicomponent cartilage relaxometry analysis.A bi-component mcDESPOT model is typically applied to characterize the fast (F) and slowly (S) relaxing water components. The mcRISE model introduces an extra macromolecule bound proton component through MT. The multicomponent T2 parameters can be obtained by fitting the model with spoiled gradient echo (SPGR) and balanced steady-state free precession (bSSFP) signals at varying flip angle (i.e. mcDESPOT) and at varying RF width (T_RF) (i.e. mcRISE). However, the assumption of instantaneous RF pulses may not be appropriate in the scenario of long RF pulse at short repetition time (TR) where the RF pulse could be around 30% of entire TR. A finite pulse correction can be performed by modification of the transverse relaxation dependent on the ratio of RF pulse width and repetition time as described as (4,6) $$$T_{2corr}=T_{2}(1-\xi\frac{T_{RF}}{TR})^{-1}$$$ where $$$\xi=0.68-0.125(1+\frac{T_{RF}}{TR})\frac{T_{2}}{T_{1}}.$$$ An incremental fitting approach is applied to fit for the parameters of mcRISE model. The bound pool fraction f and exchange rate k are first derived using varying RF pulse width bSSFP (bSSFPv) in a simplified two-pool model. Then, f and k as fixed parameters are applied in the full model to correct MT effect for T_2S, T_2F and fast water fraction f_F (Figure 1). To correct the finite pulse effect, the apparent T2 transverse time in model equations is substituted with the modified T2 for each image voxel at both simplified and full model fit.A knee scan was performed on a healthy adult using a GE MR750 3.0T scanner. The imaging parameters included: 1) SPGR scans with TR/TE=4.6/2.2ms over 8 flip angles from 3° to 18°); 2) Two bSSFP scans with RF phase cycling on (bSSFP₁₈₀) and off (bSSFP₀), with TR/TE=5.0/2.4ms over 8 flip angles from 2° to 50°); 3) One bSSFPv scans over a range of 8 RF pulse width from T_RF/TR=0.2/5.6ms to 2/6.5ms and α=35°. 4) Inversion recovery SPGR scan with TR/TE=4.6/2.2ms, TI=450ms, and α=5°. All scans were performed in the sagittal plane with a 16cm field of view, 256×256 matrix, 3mm thickness. The full set of image data was applied to mcRISE model fit with and without finite pulse correction. A subset of image data excluding bSSFPv was applied to the mcDEPOT model fit. Relaxometry parameters including T₂, T_2F, T_2S, f_F, f and k are compared across different methods.mcRISE was able to create T2 and water fraction maps for the both fast and slow relaxing water components and f and k maps of the entire knee joint at 3.0T (Figure 2). As shown in our previous studies, mcRISE provides higher single component T₂, T_2F, T_2S and lower f_F values compared to values obtained from mcDESPOT (Table 1)(3). The mcDESPOT parameters are strongly influenced by the MT exchange, correcting which can provide MT in-sensitive multicomponent T2 parameters analysis. The finite pulse effect has a non-negligible influence on the steady-state sequences. The T₂, f and k are lower in the mcRISE after finite pulse correction which agrees well with previous studies using finite pulse correction in the brain imaging (7). In addition, the T_2F and T_2S have lower values in the mcRISE after finite pulse correction, while the fast water faction f_F shows an increased value after correction. This in-vivo finding corresponds well with our simulation data (not shown). Due to the fact that the finite pulse and MT effect could influence multi-component T2 parameters in a rather complex way, this proposed combined finite pulse and MT correction may provide more robust and accurate relaxometry analysis of articular cartilage. However, there is need for further studies in human subjects and ex-vivo studies on cartilage specimens with histologic correlation to determine the sensitivity and specificity of these new parameters to disease-related changes in the cartilage matrix.